RE: IDS 10.0 performance on RedHat

Posted in 2006

Alexey Sonkin — Wed, 19 Jul 2006 19:59:15 -0400 — source: Usenet: comp.databases.informix

Neil, I think, that both data load (into unindexed table) and index build (which is actually a sequential scan from I/O standpoint) are more hungry for latency, then for parallelizm. Also, it is quite clear, why increasing number of CPU VP's impoves KAIO performance: IDS directly uses cpu vp's to 'poll' kaio request for completion, one kiao thread per cpu vp. The more vp's, the more kaio threads, the more throughput. And, still, try another test, like index join with PDQPRIORITY, and compare results. ------------------------------------------ Alexey Sonkin -----Original Message----- From: informix-list-bounces@iiug.org [mailto:informix-list-bounces@iiug.org] On Behalf Of Neil Truby Sent: Wednesday, July 19, 2006 7:09 PM To: informix-list@iiug.org Subject: Re: IDS 10.0 performance on RedHat >> My observation is that for parallel load, Informix AIO gives up to 100% better throughput (I don't use term 'performance' here!), then AIO. Just allocate as many AIO VP's, as you need (30+)! Unless I'm mis-understanding you, your results are not similar to mine. With low numbers of cpu vps I found kaio slower than using aio, but with as many or more cpus as cores, that is >= 8, I found kaio superior, even against a test with 127 aio vps. _______________________________________________ Informix-list mailing list Informix-list@iiug.org http://www.iiug.org/mailman/listinfo/informix-list